MAIN FIGURES
col_fun <-
colorRamp2(
colors = c(col_a,"#eeeeee",col_b),
breaks = c(.75,1,1.25)
)
lgd = Legend(col_fun = col_fun, title = "FC", at = c(.75, 1.25), labels = c("EB", "LG"))
par(mar = c(5,4,4,3)+.1)
plot(
main = "Changes in\nTF centrality",
x = relativise(EB_vs_LG$centrality_changes_TFs$a),
y = relativise(EB_vs_LG$centrality_changes_TFs$b),
pch = 21,
col = darken(col_fun(EB_vs_LG$centrality_changes_TFs$b/EB_vs_LG$centrality_changes_TFs$a),.5),
bg = alpha(col_fun(EB_vs_LG$centrality_changes_TFs$b/EB_vs_LG$centrality_changes_TFs$a),.8),
xlab = "Early Blastula",
ylab = "Late Gastrula",
bty = "n"
)
abline(a = 0, b = 1, col = "#4c4c4c", lty = 2)
draw(lgd, x = unit(.92, "npc"), y = unit(.5, "npc"))
par(mar = c(5,4,4,2)+.1)
PDF:
pdf("graphics/3E.pdf", wi = 3.5, he = 3.8)
par(mar = c(5,4,4,3)+.1)
plot(
main = "Changes in\nTF centrality",
x = relativise(EB_vs_LG$centrality_changes_TFs$a),
y = relativise(EB_vs_LG$centrality_changes_TFs$b),
pch = 21,
col = darken(col_fun(EB_vs_LG$centrality_changes_TFs$b/EB_vs_LG$centrality_changes_TFs$a),.5),
bg = alpha(col_fun(EB_vs_LG$centrality_changes_TFs$b/EB_vs_LG$centrality_changes_TFs$a),.8),
xlab = "Early Blastula",
ylab = "Late Gastrula",
bty = "n"
)
abline(a = 0, b = 1, col = "#4c4c4c", lty = 2)
draw(lgd, x = unit(.92, "npc"), y = unit(.5, "npc"))
par(mar = c(5,4,4,2)+.1)
dev.off()
## png
## 2
Centrality per TF class:
pfla_tfs_graph_analysis <- pfla_attributes_list[[2]]
allTFclasses_col <- c(topclasses_col,otherclasses_col)
## Centrality per TF class
pfla_tfs_graph_analysis$TFclass <- factor(pfla_tfs_graph_analysis$TFclass, levels = unique(sort(pfla_tfs_graph_analysis$TFclass)))
EB_vs_LG$centrality_TFclass_a$class <- factor(EB_vs_LG$centrality_TFclass_a$class, levels = levels(pfla_tfs_graph_analysis$TFclass))
EB_vs_LG$centrality_TFclass_b$class <- factor(EB_vs_LG$centrality_TFclass_b$class, levels = levels(pfla_tfs_graph_analysis$TFclass))
allTFclasses_col <- allTFclasses_col[match(levels(pfla_tfs_graph_analysis$TFclass), names(allTFclasses_col))]
class_cent_plot <- as_ggplot( as.grob( function(){
par(mfrow = c(2,1))
set.seed(4344)
graphics::boxplot(
main = "Early Blastula",
relativise(EB_vs_LG$centrality_TFclass_a$centr)~EB_vs_LG$centrality_TFclass_a$class,
ylab = "relative centrality",
xlab = "",
col = allTFclasses_col, border = colorspace::darken(allTFclasses_col, .5),
las = 2,
cex.axis = .8,
outline = F
)
stripchart(
relativise(EB_vs_LG$centrality_TFclass_a$centr)~EB_vs_LG$centrality_TFclass_a$class,
col = colorspace::darken(allTFclasses_col, .6),
method = "jitter",
jitter=0.15,
vertical = TRUE,
pch = 20,
cex=0.8,
add = TRUE
)
set.seed(4344)
graphics::boxplot(
main = "Late Gastrula",
relativise(EB_vs_LG$centrality_TFclass_b$centr)~EB_vs_LG$centrality_TFclass_b$class,
ylab = "relative centrality",
xlab = "",
col = allTFclasses_col, border = colorspace::darken(allTFclasses_col, .5),
las = 2,
cex.axis = .8,
outline = F
)
stripchart(
relativise(EB_vs_LG$centrality_TFclass_b$centr)~EB_vs_LG$centrality_TFclass_b$class,
col = colorspace::darken(allTFclasses_col, .6),
method = "jitter",
jitter=0.15,
vertical = TRUE,
pch = 20,
cex=0.8,
add = TRUE
)
par(mfrow = c(1,1))
}))
grid.draw(class_cent_plot)
pdf("graphics/3F.pdf", he = 8, wi = 12)
grid.draw(class_cent_plot)
dev.off()
## png
## 2
Influence plot:
contains <- function(x,a){
y = unlist(sapply(x,function(x)length(grep(a,x)) != 0))
return(y)
}
x <- EB_vs_LG$influence_results$influence_table
x$genename <-
translate_ids(
x = x$factor,
dict = insitu_data[,c(1,4)]
)
x$genename[contains(x$genename,"TCONS")] <-
paste0(
translate_ids(
x$factor[contains(x$genename,"TCONS")],
pfla_genenames[,c(1,2)]
),
"-like"
)
x$genename[contains(x$genename,"TCONS")] <-
x$TFclass[contains(x$genename,"TCONS")]
par(mar = c(5,4,4,8)+.1, xpd = T)
plot(
main="Main factors",
x$factor_fc, x$influence_score,
pch = 21, bg = x$col, col = darken(x$col,.5),
xlab="log2fold change of TF",
ylab="ANANSE influence score",
bty="n",
cex = 1.1,
xlim= c(0,max(x$factor_fc) + 1)
)
text(
x = x$factor_fc[1:20],
y = x$influence_score[1:20],
labels = x$genename[1:20],
pos = 4,
cex = .7
)
legend(
"bottomright",
legend = unique(x$TFclass),
pch = 21,
inset=c(-0.3, 0),
col = darken(unique(x$col),.5),
pt.bg = unique(x$col,.5),
cex = .7,
ncol = 1
)
par(mar = c(5,4,4,2)+.1)
PDF:
pdf("graphics/3E.pdf", he = 5, wi = 6)
par(mar = c(5,4,4,8)+.1, xpd = T)
plot(
main="Main factors",
x$factor_fc, x$influence_score,
pch = 21, bg = x$col, col = darken(x$col,.5),
xlab="log2fold change of TF",
ylab="ANANSE influence score",
bty="n",
cex = 1.1,
xlim= c(0,max(x$factor_fc) + 1)
)
text(
x = x$factor_fc[1:20],
y = x$influence_score[1:20],
labels = x$genename[1:20],
pos = 4,
cex = .7
)
legend(
"bottomright",
legend = unique(x$TFclass),
pch = 21,
inset=c(-0.3, 0),
col = darken(unique(x$col),.5),
pt.bg = unique(x$col,.5),
cex = .7,
ncol = 1
)
par(mar = c(5,4,4,2)+.1)
dev.off()
## png
## 2
SUPPLEMENTARY FIGURES
S14
par(mfrow = c(1,5))
barplot(
c(
EB_vs_LG$comparative$a[1],
EB_vs_LG$comparative$b[1]
),
main = EB_vs_LG$comparative$metric[1],
names.arg = c("EB","LG"),
col = c(col_a,col_b),
border = darken(c(col_a,col_b),.5),
ylim = c(0,15000)
)
barplot(
c(
EB_vs_LG$comparative$a[2],
EB_vs_LG$comparative$b[2]
),
main = EB_vs_LG$comparative$metric[2],
col = c(col_a,col_b),
border = darken(c(col_a,col_b),.5),
ylim = c(0,350),
names.arg = c("EB","LG")
)
boxplot(
main = "connections per\ntarget gene",
list(
V(g_EB)$indegree[V(g_EB)],
V(g_LG)$indegree[V(g_LG)]
),
names = c("EB", "LG"),
col = c(col_a,col_b),
border = darken(c(col_a,col_b),.5),
sub = paste0(
"Wilcox p.value ",
wilcox.test(
x = V(g_EB)$indegree[V(g_EB)],
y = V(g_LG)$indegree[V(g_LG)]
)$p.value
)
)
boxplot(
main = "relative outdegree",
list(
V(g_EB)$rel_outdegree[V(g_EB)$is_TF],
V(g_LG)$rel_outdegree[V(g_LG)$is_TF]
),
names = c("EB", "LG"),
col = c(col_a,col_b),
border = darken(c(col_a,col_b),.5),
sub = paste0(
"Wilcox p.value ",
wilcox.test(
V(g_EB)$rel_outdegree[V(g_EB)$is_TF],
V(g_LG)$rel_outdegree[V(g_LG)$is_TF]
)$p.value
)
)
# Number of self-reg TFs
barplot(
c(
EB_vs_LG$comparative$a[5],
EB_vs_LG$comparative$b[5]
),
main = EB_vs_LG$comparative$metric[5],
col = c(col_a,col_b),
border = darken(c(col_a,col_b),.5),
names.arg = c("EB","LG")
)
par(mfrow = c(1,1))
pdf("graphics/S14_1.pdf", he = 3, wi = 8)
{
par(mfrow = c(1,5))
barplot(
c(
EB_vs_LG$comparative$a[1],
EB_vs_LG$comparative$b[1]
),
main = EB_vs_LG$comparative$metric[1],
names.arg = c("EB","LG"),
col = c(col_a,col_b),
border = darken(c(col_a,col_b),.5),
ylim = c(0,15000)
)
barplot(
c(
EB_vs_LG$comparative$a[2],
EB_vs_LG$comparative$b[2]
),
main = EB_vs_LG$comparative$metric[2],
col = c(col_a,col_b),
border = darken(c(col_a,col_b),.5),
ylim = c(0,350),
names.arg = c("EB","LG")
)
boxplot(
main = "connections per\ntarget gene",
list(
V(g_EB)$indegree[V(g_EB)],
V(g_LG)$indegree[V(g_LG)]
),
names = c("EB", "LG"),
col = c(col_a,col_b),
border = darken(c(col_a,col_b),.5),
sub = paste0(
"Wilcox p.value ",
wilcox.test(
x = V(g_EB)$indegree[V(g_EB)],
y = V(g_LG)$indegree[V(g_LG)]
)$p.value
)
)
boxplot(
main = "relative outdegree",
list(
V(g_EB)$rel_outdegree[V(g_EB)$is_TF],
V(g_LG)$rel_outdegree[V(g_LG)$is_TF]
),
names = c("EB", "LG"),
col = c(col_a,col_b),
border = darken(c(col_a,col_b),.5),
sub = paste0(
"Wilcox p.value ",
wilcox.test(
V(g_EB)$rel_outdegree[V(g_EB)$is_TF],
V(g_LG)$rel_outdegree[V(g_LG)$is_TF]
)$p.value
)
)
# Number of self-reg TFs
barplot(
c(
EB_vs_LG$comparative$a[5],
EB_vs_LG$comparative$b[5]
),
main = EB_vs_LG$comparative$metric[5],
col = c(col_a,col_b),
border = darken(c(col_a,col_b),.5),
names.arg = c("EB","LG")
)
par(mfrow = c(1,1))
}
dev.off()
## png
## 2
col_fun <-
colorRamp2(
colors = c(col_a,"#eeeeee",col_b),
breaks = c(.8,1,1.2)
)
lgd = Legend(col_fun = col_fun, title = "FC", at = c(.8, 1.2), labels = c("EB", "LG"))
par(mar = c(5,4,4,3)+.1)
plot(
main = "Changes in\ntrans-dev centrality",
x = relativise(EB_vs_LG$centrality_changes_transdev$a),
y = relativise(EB_vs_LG$centrality_changes_transdev$b),
col = darken(col_fun(EB_vs_LG$centrality_changes_transdev$b/EB_vs_LG$centrality_changes_transdev$a),.5),
bg = alpha(col_fun(EB_vs_LG$centrality_changes_transdev$b/EB_vs_LG$centrality_changes_transdev$a),.8),
xlab="Early Blastula",
ylab = "Late Gastrula",
pch = 21,
bty = "n"
)
abline(a = 0, b = 1, col = "#4c4c4c", lty = 2)
draw(lgd, x = unit(.92, "npc"), y = unit(.5, "npc"))
par(mar = c(5,4,4,2)+.1)
pdf("graphics/S14_2.pdf", wi = 3.5, he = 3.8)
{
par(mar = c(5,4,4,3)+.1)
plot(
main = "Changes in\ntrans-dev centrality",
x = relativise(EB_vs_LG$centrality_changes_transdev$a),
y = relativise(EB_vs_LG$centrality_changes_transdev$b),
col = darken(col_fun(EB_vs_LG$centrality_changes_transdev$b/EB_vs_LG$centrality_changes_transdev$a),.5),
bg = alpha(col_fun(EB_vs_LG$centrality_changes_transdev$b/EB_vs_LG$centrality_changes_transdev$a),.8),
xlab="Early Blastula",
ylab = "Late Gastrula",
pch = 21,
bty = "n"
)
abline(a = 0, b = 1, col = "#4c4c4c", lty = 2)
draw(lgd, x = unit(.92, "npc"), y = unit(.5, "npc"))
par(mar = c(5,4,4,2)+.1)
}
dev.off()
## png
## 2
S15
GO_TFs <-
plot_grid(
EB_vs_LG$GOs_TFs$GOplot$TFs_exclusive_a,
EB_vs_LG$GOs_TFs$GOplot$TFs_common_ab,
EB_vs_LG$GOs_TFs$GOplot$TFs_exclusive_b,
ncol = 1
)
GO_TFs
pdf("graphics/S15_1.pdf", wi = 5, he = 12)
GO_TFs
dev.off
## function (which = dev.cur())
## {
## if (which == 1)
## stop("cannot shut down device 1 (the null device)")
## .External(C_devoff, as.integer(which))
## dev.cur()
## }
## <bytecode: 0x4147158>
## <environment: namespace:grDevices>
GO_TGs <-
plot_grid(
EB_vs_LG$GOs_targets$GOplot$tgs_exclusive_a_top,
EB_vs_LG$GOs_targets$GOplot$tgs_common_ab_top,
EB_vs_LG$GOs_targets$GOplot$tgs_exclusive_b_top,
ncol = 1
)
GO_TGs
pdf("graphics/S15_2.pdf", wi = 5, he = 12)
GO_TGs
dev.off()
## png
## 2
fit_tg <- euler(calc_overlaps(EB_vs_LG$tg_overlap_top))
cols_venn <- c(col_a,col_b)
plot(
fit_tg,
fills = alpha(cols_venn,0.45),
edges = darken(cols_venn,.5),
quantities = list(type = c("counts", "percent"))
)
pdf("graphics/S15_3.pdf", wi = 4, he = 2)
plot(
fit_tg,
fills = alpha(cols_venn,0.45),
edges = darken(cols_venn,.5),
quantities = list(type = c("counts", "percent"))
)
dev.off()
## png
## 2
fit_tf <- euler(calc_overlaps(EB_vs_LG$TF_overlap))
cols_venn <- c(col_a,col_b)
plot(
fit_tf,
fills = alpha(cols_venn,0.45),
edges = darken(cols_venn,.5),
quantities = list(type = c("counts", "percent"))
)
pdf("graphics/S15_4.pdf", wi = 4, he = 2)
plot(
fit_tf,
fills = alpha(cols_venn,0.45),
edges = darken(cols_venn,.5),
quantities = list(type = c("counts", "percent"))
)
dev.off()
## png
## 2
pfla_funcat <- pfla_attributes_list[[3]]
funcat_lgd =
Legend(
labels = paste0(LETTERS,") ",funcat_lookup$functional_category[match(LETTERS,funcat_lookup$cog)]),
title = "Functional Categories",
legend_gp = gpar(
fill = colorspace::lighten(desaturate(rainbow(26),.3))
),
border = colorspace::darken(desaturate(rainbow(26),.3),.5)
)
funcat <- as_ggplot( as.grob( function(){
par(mfrow = c(2,1))
par(mar = c(5,4,4,16)+.1)
barplot(
main = "EB",
log1p(table(factor(pfla_funcat$funcat, levels = LETTERS)[
pfla_funcat$id %in% V(g_EB)$name[V(g_EB)$rel_outdegree > 0.5]
] )),
ylab = "log1p(no. TFs)",
col = colorspace::lighten(desaturate(rainbow(26),.3)),
border = colorspace::darken(desaturate(rainbow(26),.3),.5)
)
barplot(
main = "LG",
log1p(table(factor(pfla_funcat$funcat, levels = LETTERS)[
pfla_funcat$id %in% V(g_LG)$name[V(g_LG)$rel_outdegree > 0.5]
] )),
ylab = "log1p(no. TFs)",
col = colorspace::lighten(desaturate(rainbow(26),.3)),
border = colorspace::darken(desaturate(rainbow(26),.3),.5)
)
# draw(funcat_lgd, x = unit(.9, "npc"), y = unit(.5, "npc"))
par(mar = c(5,4,4,2)+.1)
par(mfrow = c(1,1))
} ))
grid.draw(funcat)
draw(funcat_lgd, x = unit(.8, "npc"), y = unit(.5, "npc"))
pdf("graphics/S15_5.pdf", he = 7, wi = 10)
grid.draw(funcat)
draw(funcat_lgd, x = unit(.8, "npc"), y = unit(.5, "npc"))
dev.off()
## png
## 2
# in EB
chisq_cog_EB_tgs <-
make_chisq_heatmap(chisq_and_posthoc(
table(
pfla_funcat$funcat,
ifelse(pfla_funcat$id %in% EB_vs_LG$tg_overlap$tgs_exclusive_a, "EB_tg", "non_EB_tg")
)
))
# in LG
chisq_cog_LG_tgs <-
make_chisq_heatmap(chisq_and_posthoc(
table(
pfla_funcat$funcat,
ifelse(pfla_funcat$id %in% EB_vs_LG$tg_overlap$tgs_exclusive_b, "LG_tg", "non_LG_tg")
)
))
# common
chisq_cog_common_tgs <-
make_chisq_heatmap(chisq_and_posthoc(
table(
pfla_funcat$funcat,
ifelse(pfla_funcat$id %in% EB_vs_LG$tg_overlap$tgs_common_ab, "common_tg", "non_common_tg")
)
))
# Internal comparison between themselves
pfla_cogs_tg_internal <-
merge(
pfla_funcat,
data.frame(
id = unlist(EB_vs_LG$tg_overlap),
tg =rep(names(EB_vs_LG$tg_overlap), times = sapply(EB_vs_LG$tg_overlap,length))
),
by.x = 1, by.y = 1
)
chisq_cog_EB_tgs_internal <-
make_chisq_heatmap(chisq_and_posthoc(
table(
pfla_cogs_tg_internal$funcat,
pfla_cogs_tg_internal$tg
)
))
# plots
draw(
chisq_cog_EB_tgs$heatmap %v%
chisq_cog_common_tgs$heatmap %v%
chisq_cog_LG_tgs$heatmap %v%
chisq_cog_EB_tgs_internal$heatmap
)
pdf("graphics/S15_6.pdf", he = 4.5, wi = 8)
draw(
chisq_cog_EB_tgs$heatmap %v%
chisq_cog_common_tgs$heatmap %v%
chisq_cog_LG_tgs$heatmap %v%
chisq_cog_EB_tgs_internal$heatmap
)
dev.off()
## png
## 2
# number of post-gastrulation genes in network
barplot(
main="network-specific targets\nin post-gastrulation clusters",
height=c(
EB_vs_LG$comparative$a[15]/length(EB_vs_LG$tg_overlap$tgs_exclusive_a)*100,
EB_vs_LG$comparative$b[15]/length(EB_vs_LG$tg_overlap$tgs_exclusive_b)*100
),
col = c(col_a,col_b),
border = darken(c(col_a,col_b),.4),
names=c("EB", "LG"),
ylab="gene percent",
ylim = c(0,100),
cex.main = .7,
las=1
)
pdf("graphics/S15_6.pdf", wi = 2, he = 3.5)
barplot(
main="network-specific targets\nin post-gastrulation clusters",
height=c(
EB_vs_LG$comparative$a[15]/length(EB_vs_LG$tg_overlap$tgs_exclusive_a)*100,
EB_vs_LG$comparative$b[15]/length(EB_vs_LG$tg_overlap$tgs_exclusive_b)*100
),
col = c(col_a,col_b),
border = darken(c(col_a,col_b),.4),
names=c("EB", "LG"),
ylab="gene percent",
ylim = c(0,100),
cex.main = .7,
las=1
)
dev.off()
## png
## 2
S16
ecto_genes <- which(
V(g_LG)$germlayer== "Ectoderm")
meso_genes <- which(
V(g_LG)$germlayer== "Mesoderm" |
V(g_LG)$germlayer=="EctoMeso" |
V(g_LG)$germlayer=="EndoMeso" |
V(g_LG)$germlayer== "All"
)
endo_genes <- which(
V(g_LG)$germlayer== "Endoderm" |
V(g_LG)$germlayer=="EndoEcto" |
V(g_LG)$germlayer=="EndoMeso" |
V(g_LG)$germlayer== "All"
)
#ectoderm
ecto_graph <- induced_subgraph(
g_LG,
vids = ecto_genes,
impl = "auto"
)
ecto_graph <- subgraph_with_top_edges_per_tg(ecto_graph,top = 2,mode="both")
V(ecto_graph)$genename <- translate_ids(x = V(ecto_graph)$name , dict = insitu_data[,c(1,4)])
ecto_graph$weight <- cut(E(ecto_graph)$width,breaks = quantile(E(ecto_graph)$width), include.lowest = TRUE, labels = FALSE)
V(ecto_graph)$col[V(ecto_graph)$col == ""] <- "#dddddd"
#mesoderm
meso_graph <- induced_subgraph(
g_LG,
vids = meso_genes,
impl = "auto"
)
meso_graph <- subgraph_with_top_edges_per_tg(meso_graph, top = 2,mode="both")
V(meso_graph)$genename <- translate_ids(x = V(meso_graph)$name , dict = insitu_data[,c(1,4)])
meso_graph$weight <- cut(E(meso_graph)$width,breaks = quantile(E(meso_graph)$width), include.lowest = TRUE, labels = FALSE)
V(meso_graph)$col[V(meso_graph)$col == ""] <- "#dddddd"
#endoderm
endo_graph <- induced_subgraph(
g_LG,
vids = endo_genes,
impl = "auto"
)
endo_graph <- subgraph_with_top_edges_per_tg(endo_graph,top = 2,mode="both")
V(endo_graph)$genename <- translate_ids(x = V(endo_graph)$name , dict = insitu_data[,c(1,4)])
endo_graph$weight <- cut(E(endo_graph)$width,breaks = quantile(E(endo_graph)$width), include.lowest = TRUE, labels = FALSE)
V(endo_graph)$col[V(endo_graph)$col == ""] <- "#dddddd"
germlayer_graphs <- as_ggplot( as_grob( function(){
par(mfrow = c(3,1))
#ectoderm
set.seed(1234)
plot(
main = "Ectoderm graph",
ecto_graph,
vertex.color = V(ecto_graph)$germlayer_col,
vertex.frame.color = darken(V(ecto_graph)$germlayer_col,.5),
vertex.label = V(ecto_graph)$genename,
vertex.size = 8,
vertex.label.family = "Helvetica",
vertex.label.color = "black",
edge.width = ecto_graph$weight,
vertex.label.cex = .9,
edge.arrow.size = 0.2,
edge.color = rgb(0,0,0,0.1),
layout = layout.graphopt(ecto_graph)
)
#mesoderm
set.seed(1234)
plot(
main = "Mesoderm graph",
meso_graph,
vertex.color = V(meso_graph)$germlayer_col,
vertex.frame.color = darken(V(meso_graph)$germlayer_col,.5),
vertex.label = V(meso_graph)$genename,
vertex.size = 8,
vertex.label.family = "Helvetica",
vertex.label.color = "black",
vertex.label.cex = .9,
edge.width = meso_graph$weight,
edge.arrow.size = 0.2,
edge.color = rgb(0,0,0,0.1),
layout = layout.graphopt(meso_graph)
)
#endoderm
set.seed(1234)
plot(
main = "Endoderm graph",
endo_graph,
vertex.color = V(endo_graph)$germlayer_col,
vertex.frame.color = darken(V(endo_graph)$germlayer_col,.5),
vertex.label = V(endo_graph)$genename,
vertex.size = 8,
vertex.label.family = "Helvetica",
vertex.label.color = "black",
vertex.label.cex = .9,
edge.width = endo_graph$weight,
edge.arrow.size = 0.2,
edge.color = rgb(0,0,0,0.1),
layout = layout.graphopt(endo_graph)
)
par(mfrow = c(1,1))
}))
grid.draw(germlayer_graphs)
pdf("graphics/S16_1.pdf", wi = 5, he = 15)
grid.draw(germlayer_graphs)
dev.off()
## png
## 2
cent_germlyr <-
as.data.frame(vertex_attr(g_LG, index = V(g_LG)[[which(V(g_LG)$germlayer != "")]]))
cent_germlyr$eggNOG.annot <- NULL
cent_germlyr$layer <- cent_germlyr$germlayer
cent_germlyr$layer[cent_germlyr$germlayer == "EctoMeso"] = "Mesoderm" # anything that is in mesoderm
cent_germlyr$layer[cent_germlyr$germlayer == "EndoMeso"] = "Endoderm" # anything that is in endoderm
cent_germlyr$layer[cent_germlyr$germlayer == "EndoEcto"] = "Endoderm" # anything that is in endoderm
cent_germlyr$layer =
factor(cent_germlyr$layer, levels = c("Ectoderm","Mesoderm","Endoderm", "All"))
lyr_cols =
c(
Ectoderm = "#3d85c6",
Mesoderm = "#e26767",
Endoderm = "#fcd086",
All = "#dddddd"
)
set.seed(1234)
p_cent_germlyr <-
cent_germlyr %>% group_by(layer) %>%
mutate(rel_centr = relativise(centr)) %>%
ggplot(aes(
y = layer, x = rel_centr,
colour = layer, fill = layer, label = gene
))+
geom_jitter(height = 0.2, size = 3, shape = 21)+
scale_y_discrete(limits = rev(levels(cent_germlyr$layer)))+
scale_fill_manual(values = lyr_cols)+
scale_colour_manual(values = darken(lyr_cols,.4))+
theme_classic()+
xlab("Relative centrality")+
ylab("Germ Layer")+
geom_text_repel(show_guide = FALSE)
print(p_cent_germlyr)
pdf("graphics/S16_2.pdf", wi = 6, he = 4.5)
print(p_cent_germlyr)
dev.off()
## png
## 2
tfs_centr <- EB_vs_LG$centrality_TFclass_b
plot(density(tfs_centr$centr))
abline(v=quantile(tfs_centr$centr, c(0, .33, .67, 1)), col = desaturate(lighten(rainbow(4),.4),.4), lwd = 2)
qs <- cut(tfs_centr$centr, breaks = quantile(tfs_centr$centr, c(0, 0.33, 0.67, 1)), labels = FALSE, include.lowest = TRUE)
grouped_tfs <- split(tfs_centr$id, qs)
names(grouped_tfs) <- c("low_central","mid_central","high_central")
gos_by_tfcentr <- getGOs(
genelist = grouped_tfs,
gene_universe = tfs_centr$id,
alg = "elim",
gene2GO = topGO::readMappings("outputs/functional_annotation/go_blast2go/GO_annotation.txt")
)
## [1] "Starting analysis 1 of 3"
## [1] "Starting analysis 2 of 3"
## [1] "Starting analysis 3 of 3"
plot_grid(
gos_by_tfcentr$GOplot$low_central,
gos_by_tfcentr$GOplot$mid_central,
gos_by_tfcentr$GOplot$high_central,
ncol = 1
)
pdf("graphics/S16_3.pdf", he = 12, wi = 6)
plot_grid(
gos_by_tfcentr$GOplot$low_central,
gos_by_tfcentr$GOplot$mid_central,
gos_by_tfcentr$GOplot$high_central,
ncol = 1
)
dev.off()
## png
## 2
g_inf <- EB_vs_LG$influence_results$influence_subgraph
V(g_inf)$genename <-
translate_ids(V(g_inf)$name,pfla_genenames)
E(g_inf)$width <-
category_by_quantile(
E(g_inf)$prob,
newvalues = c(0.2,1,2,3.5)
)
V(g_inf)$col[V(g_inf)$col == ""] <- "gray"
V(g_inf)$genename <- gsub("TCONS_","", V(g_inf)$genename)
V(g_inf)$genename[V(g_inf)$genename == "T"] <- "BRA"
V(g_inf)$genename[V(g_inf)$name %in% insitu_data$id] <-
translate_ids(
x = V(g_inf)$name[V(g_inf)$name %in% insitu_data$id],
dict = insitu_data[,c(1,4)]
)
V(g_inf)$cex <- as.numeric(cut(V(g_inf)$size, breaks = quantile(V(g_inf)$size,c(0,.5,1)), labels = c(1,1.75), include.lowest = TRUE))
E(g_inf)$color <- tail_of(g_inf,E(g_inf))$col
set.seed(1234)
plot(
main = "Influence network",
g_inf,
vertex.size = 5*V(g_inf)$cex,
vertex.color = V(g_inf)$col,
vertex.label.family = "Helvetica",
vertex.label.color = "black",
vertex.label.cex = .75,
vertex.label.dist = 1,
vertex.label = V(g_inf)$genename,
edge.width = E(g_inf)$width,
edge.arrow.size = 0.3,
edge.color = alpha(E(g_inf)$color, .6),
layout = layout_with_dh(g_inf),
edge.curved = .1
)
pdf("graphics/S16_4.pdf",he = 6.5, wi = 6.5)
set.seed(1234)
plot(
main = "Influence network",
g_inf,
vertex.size = 5*V(g_inf)$cex,
vertex.color = V(g_inf)$col,
vertex.label.family = "Helvetica",
vertex.label.color = "black",
vertex.label.cex = .75,
vertex.label.dist = 1,
vertex.label = V(g_inf)$genename,
edge.width = E(g_inf)$width,
edge.arrow.size = 0.3,
edge.color = alpha(E(g_inf)$color, .6),
layout = layout_with_dh(g_inf),
edge.curved = .1
)
dev.off()
## png
## 2